Lüge - fotres

Lügendetektor -­‐ nur ein Mythos? Harald T. Schupp University of Konstanz Psychologie
Physiologie
Lüge
Galvanische
Hautreaktion
One-to-One Mapping
Cacioppo & Tassinary, 1992
Psychologie
Physiologie
Blutdruck
Lüge
Galvanische
Hautreaktion
One-to-More Mapping
Psychologie
Physiologie
Blutdruck
Galvanische
Hautreaktion
Puls
Lüge
Mikromimik
Atmung
Stimme
One-to-Many Mapping
Blutdruck
Galvanische
Hautreaktion
Haben Sie
ihre Frau getötet?
Puls
Mikromimik
Atmung
Stimme
Psychologie
Physiologie
Angst
Lüge
Galvanische
Hautreaktion
More-to-One Mapping
Psychologie
Physiologie
Anspannung
Angst
Lüge
Galvanische
Hautreaktion
Schuld
Scham
Kognitive
Belastung
Many-to-OneMapping
Psychologie
Physiologie
Anspannung
Blutdruck
Angst
Galvanische
Hautreaktion
Lüge
Puls
Erregung
Mikromimik
Scham
Atmung
Kognitive
Belastung
Stimme
Many-to-Many Mapping
Die theoreBsche PerspekBve „No known physiological response or pattern of
responses is unique to deception“ (Raskin, 1986)
Die theoreBsche PerspekBve „No known physiological response or pattern of
responses is unique to deception“ (Raskin, 1986)
„Lying is associated with increased arousal that
cannot be suppressed voluntarily“ (Fiedler et al.,
2002)
Ablauf des Vergleichsfragentests
Akteneinsicht
Biographie
Tat
Interview
Tat
Erklärung des Tests
Formulierung der Fragen
Testdurchführung
Zahlentest
Tattest (3-4 Wiederholungen)
Fragentypen
Relevante Fragen
Haben Sie ihre Frau getötet?
Vergleichs- (Kontroll-) Fragen
Haben Sie jemals jemand verprügelt?
Irrelevante Fragen
Sitzen Sie in einem Stuhl?
Antwort: Ja/Nein
Entscheidungsregeln
Lüge
Größere physiologische Reaktion für Tat- als Kontrollfragen
Wahrheit
Größere physiologische Reaktion für Kontroll- als Tatfragen
Keine Entscheidung
Inkonsistentes Reaktionsmuster
Prädiktive Validitätsstudien
Honts, 1996
‚Sampling Bias‘
„Warum die Akkuratheit überschätzt wird“
(1)  Nur Fälle werden eingeschlossen, bei denen Schuld oder Unschuld bekannt
sind
(2)  Test und Kriterium sind nicht unabhängig
Ø  Korrekt klassifiziert: Schuldige, die beim Test erkannt werden, geben später
Schuld zu
Ø  Korrekt klassifiziert: Schuld wird durch zusätzliche Fakten (bzw. Indizien)
belegt
Ø  Typischerweise nicht in die Studie eingeschlossen:
Personen, die weder ihre Schuld zugeben noch durch zusätzliche Fakten
(bzw. Indizien) nach dem Test entlastet werden
Fiedler et al., 2002
Iacono, 2008
‚Sampling Bias‘
„Warum die Akkuratheit überschätzt wird“
Nur Fälle werden eingeschlossen, bei denen Schuld oder Unschuld bekannt sind
Ø  Korrekt klassifiziert: Schuldige, die beim Test erkannt werden, geben später
Schuld zu
Ø  Korrekt klassifiziert: Schuld wird durch zusätzliche Fakten (bzw. Indizien)
belegt
Ø  Typischerweise nicht in die Studie eingeschlossen:
Personen, die weder ihre Schuld zugeben noch durch zusätzliche Fakten
(bzw. Indizien) nach dem Test entlastet werden
„...,the best available field studies support accuracies of 57 and 75 % for
innocent and guilty subjects.“ (Iacono, 2008)
Fiedler et al., 2002
Iacono, 2008
to be an empirical question that could be checked relatively easily through research
and experimentation, but in fact matters are more complicated and it turns out that the
research conducted to date cannot provide a simple, clear-cut answer to the question
of the CQT’s validity. In order to allow conclusions about the value of the CQT,
as typically conducted in real-life conditions, an experiment should fulfil the
following requirements (see Ben-Shakhar & Furedy, 1990; Ginton, Daie, Elaad, &
Ben-Shakhar, 1982):
Die ideale Studie
(1)
(2)
(3)
(4)
The existence of a clear, conclusive and irrefutable criterion for the guilt
or innocence of the research participants. Clearly, without such a criterion there is
no way to determine whether the CQT interrogator was right or wrong in a
particular case.
A representative sampling of examinees and of the situations in which CQTs are
employed.
Independence between the criterion and the polygraph examiner’s judgment
(which may be affected by all the information at his disposal).
Testing conditions in the experiment, which resemble those of a true examination.
In particular, it is important that the examinees be anxious about the
consequences of the test and take it seriously, and that the lie or the transgression
be real.
A review of the literature reveals that no existing experiments (with the possible
exception of the Ginton et al., 1982 study) meet all these requirements. In particular,
there are no experiments that simultaneously fulfil both the first and the last
requirement. All the experiments providing a satisfactory criterion are simulations
(‘mock crimes’), in which the participants know that they are participating in a role
playing game. The participants designated as ‘guilty’ are asked by the experimenter
to
Ben-Shakkar, 2008
steal an envelope containing money, or some other item. Then, all the participants (both
decision model, this spike represents the cut point for the discrimination between lies and true answers to the critical questions. Is this model assumption justified? This central psychological problem amounts to pursuing the following concrete
questions: Is there a general law in emotion psychology, or in
psychophysiology, stating that unidimensional arousal measures can be used to identify internal mental states of consciousness, or affective states (such as lying)? Is the assumption justified that control and critical questions constitute
separable constructs, or arrow spikes in the figure? How does
Die Bedeutung der Kontrollfragen
Reaction to
control questions
Aggregate
Intensity
Truth
Lie
Orienting reactions
Fear, embarrassment, disgust
Countermeasures: biting tongue
Threat of test situation
FIGURE 1 The implicit reaction model, or decision model, underlying the Control-Question Test.
limit for classifying "criminal" lies? Systematic r
this problem, that is sensitive to the modem state
should at least refer to research on question select
tifacts and research on autobiographical memory
terviewing and survey methodology (Schwarz &
1994). If CQT research has forgotten to take thes
ately relevant research findings into account, it ca
to meet the current state of the art in scientific psy
One other crucial problem that needs to be
within the framework depicted in Figure 1 is wh
results can be manipulated successfully, using
countermeasures. Expert opinions (Iacono & Lykk
as well as empirical research findings, agree that
impossible to suppress reactions to critical que
physical as well as mental countermeasures (ton
cognitive distracters) are successful in enhancing r
control questions, thus decreasing the chances of p
results. Even CQT proponents (Honts, Raskin, &
1994) concede that, after no more than 30 min o
about 50% of respondents can manipulate the test
(counting backward by 7) or physical (biting the
pressing toes on the floor) countermeasures. More
manipulations cannot be detected by polygraph te
One suitable methodological tool to test the
sumptions illustrated in Figure 1 is signal detectio
(Swets, 1986). Referring to the distribution of
elicited by control questions as "noise" and the d
of reactions elicited by critical questions as
noise," one has a firamework for testing the assum
the CQT can really discriminate between both
questions, rather than reflecting only response bi
ternal influences. In technical terms, the q
whether control and critical questions produce
ROC curves or only different points on the s
Fiedler et al., 2002
Der Kontrollfragentest
Ø Es ist der am häufigsten angewandte Test zur
Lügendetektion
Ø Die theoretische Grundlage ist unklar
Ø Der Test genügt wissenschaftlichen Kriterien nicht
Ø Es besteht die Möglichkeit für Gegenmaßnahmen
(‚countermeasures‘)
Der Tatwissentest
Der Tatwissentest
Psychologie
Physiologie
Tatwissen/
Relevanz
Galvanische
Hautrekation
X
P300
Lüge
Der Tatwissentest
!"#
,",
Klassische Oddball-Aufgabe:
Möglichst schnelles und sicheres
Reagieren auf seltene Aufgabenreize
;2<
?@
D$=
)%&
Tag 1: Scheinverbrechen
Tag 2: Tatwissentest
•  Ein Szenario lernen
•  6 kritische Details
•  2 Szenarien: schuldig vs.
unschuldig
•  ERP-Messung: 3 Blöcke x 4
Wiederholungen:
Blauer Mantel
Phil Jenkins
Peerch Street
•  Das Szenario durchführen
6 Zielreizen
6 Testreizen
24 irrelevante Reize
•  Aufgabe: Ja/Nein-Zielreiz
Testreize
(Szenario 1)
Zielreize
der Aufgabe
Irrelevante Reize
der Aufgabe
Blue Coat
Green Hat
Brown Shoes
Red Scarf
Gray Pants
Black Gloves
Phil Jenkins
Tim Howe
Ray Snell
Neil Rand
Gene Falk
Ralph Croft
Op Cow
Op Pig
Op Horne
OpGoat
OpSheep
OpMule
Rain File
Snow File
HailFile
WindFile
Sleet File
FogFile
Sub Plans
Ship Plans
Tank Plans
Plane Plans
Bomb Plans
Gun Plans
Perch Street
Shark Street
Cod Street
Carp Street
Pike Street
Trout Street
!"##$% &'()*+,)-*./-*
3(&4%'%5#
<)9-
2-'7(&6-#
:/-;%-*,)
8-5,/&9(&+* 3(&4%'%5#
167'%7(&+*
./-0&,(-0#
12.
F-(G'6
ABC
1'3'4-56$68
6-)?$>586$68
+(<='@
1-('2$1'3'4-56
!"##
;(('3'4-56$
EB"
;(('3'4-56$68
6-)?$-5D +(<='
9(':*'562$
;(('3'4-56
78$!"##
!(8/'
ABC
1'3'4-56$68
+(<='$>586$68
6-)?@
!"#&**+,-*(0
9(':*'562$
;(('3'4-56
78$!"##
!"#$%&'()0
1-('2$1'3'4-56
!"##
Ø  Trefferquote ist oft 85 – 95 %, aber es gibt auch
Ausnahmen mit niedrigen Trefferquoten
Ø  Studien unterscheiden sich in vielen spezifischen
Details (Aufgabeninstruktion, Datenanalyse)
Ø  Eine Firma „Brain Fingerprinting Laboratories“ bietet
den Test kommerziell an
Cogn Neurodyn (2012) 6:115±154!!
!!"!
#########################################################################################$
REVI EW
.
Brain fingerprinting: a comprehensive tutorial review of detection of
concealed information with event-related brain potentials
Lawrence A. Farwell
Received: 17 March 2011 / Revised: 26 November 2011 / Accepted: 30 January 2012 / Published online: 17 February 2012
© Springer Science+Business Media B.V. 2012
DOI 10.1007/s11571-012-9192-2. Accepted for publication version ± available at: http://www.brainwavescience.com/CODY2012.pdf. Published version
available at Cognitive Neurodynamics:$http://www.springerlink.com/content/7710950336312146/$ in slightly different format.
Abstract
Brain fingerprinting (BF) detects
concealed information stored in the brain by
measuring brainwaves. A specific EEG eventrelated potential, a P300-MERMER, is elicited by
stimuli that are significant in the present context.
BF detects P300-MERMER responses to words /
pictures relevant to a crime scene, terrorist training,
bomb-making knowledge, etc.
BF detects
information by measuring cognitive information
processing. BF does not detect lies, stress, or
emotion.
BF computes a determination of
³LQIRUPDWLRQSUHVHQW´RU³LQIRUPDWLRQDEVHQW´DQGD
statistical confidence for each individual
determination. Laboratory and field tests at the
FBI, CIA, US Navy and elsewhere have resulted in
0% errors: no false positives and no false negatives.
100% of determinations made were correct. 3% of
UHVXOWV KDYH EHHQ ³LQGHWHUPLQDWH´ %F has been
Keywords brain fingerprinting, P300-MERMER,
P300, event-related potential, detection of
concealed information
I ntroduction and background
The state of the art prior to brain fingerprinting
Brain fingerprinting is an objective, scientific
method to detect concealed information stored in
the brain by measuring electroencephalographic
(EEG) brain responses, or brainwaves, noninvasively by sensors placed on the scalp. The
technique involves presenting words, phrases, or
pictures containing salient details about a crime or
investigated situation on a computer screen, in a
series with other, irrelevant stimuli.
Brain
responses to the stimuli are measured. When the
brain processes information in specific ways,
Lawrence A. Farwell
Received: 17 March 2011 / Revised: 26 November 2011 / Accepted: 30 January 2012 / Published online: 17 February 2012
© Springer Science+Business Media B.V. 2012
DOI 10.1007/s11571-012-9192-2. Accepted for publication version ± available at: http://www.brainwavescience.com/CODY2012.pdf. P
available at Cognitive Neurodynamics:$http://www.springerlink.com/content/7710950336312146/$ in slightly different format.
Abstract
Brain fingerprinting (BF) detects
concealed information stored in the brain by
measuring brainwaves. A specific EEG eventrelated potential, a P300-MERMER, is elicited by
stimuli that are significant in the present context.
BF detects P300-MERMER responses to words /
pictures relevant to a crime scene, terrorist training,
bomb-making knowledge, etc.
BF detects
information by measuring cognitive information
processing. BF does not detect lies, stress, or
emotion.
BF computes a determination of
³LQIRUPDWLRQSUHVHQW´RU³LQIRUPDWLRQDEVHQW´DQGD
statistical confidence for each individual
determination. Laboratory and field tests at the
FBI, CIA, US Navy and elsewhere have resulted in
0% errors: no false positives and no false negatives.
100% of determinations made were correct. 3% of
UHVXOWV KDYH EHHQ ³LQGHWHUPLQDWH´ %F has been
applied in criminal cases and ruled admissible in
court.
Scientific standards for BF tests are
discussed. Meeting the BF scientific standards is
necessary for accuracy and validity. Alternative
techniques that failed to meet the BF scientific
standards produced low accuracy and susceptibility
to countermeasures. BF is highly resistant to
countermeasures. No one has beaten a BF test with
countermeasures, despite a $100,000 reward for
doing so. Principles of applying BF in the
laboratory and the field are discussed.
_____________________________________________________
L. A. Farwell
Brain Fingerprinting Laboratories, Inc., 14220 37th Ave. NE,
Keywords brain fingerprinting, P300
P300, event-related potential, d
concealed information
I ntroduction and background
The state of the art prior to brain finger
Brain fingerprinting is an objectiv
method to detect concealed informati
the brain by measuring electroence
(EEG) brain responses, or brainw
invasively by sensors placed on the
technique involves presenting words,
pictures containing salient details abou
investigated situation on a computer
series with other, irrelevant stimu
responses to the stimuli are measured
brain processes information in spe
characteristic brainwave patterns can
through computer analysis of the bra
When an individual recognizes so
significant in the current context, he ex
³$KD´UHVSRQVH7KLVUHVSRQVHLVFKD
a specific brainwave pattern known
MERMER. Brainwave responses are
determine whether or not the specific
tested is stored in the brain of the su
Brain fingerprinting computes a dete
³LQIRUPDWLRQ SUHVHQW´ ± the subject
critical infoUPDWLRQ RU ³LQIRUPDWLRQ D
does not. The system also computes
confidence for each individual determ
Psychophysiology, 41 (2004), 205–219. Blackwell Publishing Inc. Printed in the USA.
Copyright r 2004 Society for Psychophysiological Research
DOI: 10.1111/j.1469-8986.2004.00158.x
Simple, effective countermeasures to P300-based tests of
detection of concealed information
J. PETER ROSENFELD,a MATTHEW SOSKINS,a GREGORY BOSH,a and ANDREW RYANb
a
Department of Psychology, Northwestern University, Evanston, Illinois, USA
Department of Defense Polygraph Institute, Charleston, South Carolina, USA
b
Abstract
We found countermeasures to protocols using P300 in concealed information tests. One, the ‘‘six-probe’’ protocol, in
Experiment 1, uses six different crime details in one run. The countermeasure: generate covert responses to irrelevant
stimuli for each probe category. Hit rates were 82% in the guilty group; 18% in the countermeasure group. The average
reaction time (RT) distinguished these two groups, but with overlap in RT distributions. The ‘‘one-probe’’ protocol, in
the second experiment, uses one crime detail as a probe. Here, one group was run in 3 weeks as a guilty group, a
countermeasure group, and again as in Week 1. Countermeasure: Covert responses to irrelevant stimuli. In Week 1, hit
rate was 92%. In Week 2, it was 50%. In Week 3, 58%. There was no overlap in the irrelevant RT distribution in Week
2: Countermeasure use was detectable. However, in Week 3, the RT distributions resembled those of Week 1; testbeaters could not be caught. These studies have shown that tests of deception detection based on P300 amplitude as a
recognition index may be readily defeated with simple countermeasures that can be easily learned.
Descriptors: Psychophysiological detection of deception, P300, Event-related potentials, Guilty knowledge tests, Lie
detection
!"#$%&'(&)*#'&*
+ ,-.)%&/"'0&$-)-,-0%&(*
+ Auf
12&&3-!"#$"%&'"
7%0(#8#*-$)9:-.)%&/"'0&
jede der 64$%5"'%-)#6-("$")
Kategorien erfolgt
eine Aufgabe
=F?
=F?G0$H&$-I&0/&60$/&'-J&'2&9H%-)#6-KC&'*9:&$H&8-2'@9H&$--G0$H&$-I&0/&60$/&'-J&'2&9H%-)#6-KC&'*9:&$H&8-2'@9H&$--=>?
=>?G0$H&$-L0%%&860$/&'-J&'2&9H%-)#6-KC&'*9:&$H&8-2'@9H&$
G0$H&$-L0%%&860$/&'-J&'2&9H%-)#6-KC&'*9:&$H&8-2'@9H&$
=M?
=M?N)9H&8$-(0%-O'"P&(-I&:-0(-80$H&$-79:#:
N)9H&8$-(0%-O'"P&(-I&:-0(-80$H&$-79:#:
=Q?
=Q?N)9H&8$-(0%-/'"P&(-I&:-0(-'&9:%&$-79:#:
N)9H&8$-(0%-/'"P&(-I&:-0(-'&9:%&$-79:#:
=R?
=R?S"'*%&88&$B-S&'*#9:*8&0%&'-/0C%-K:'6&0/&
S"'*%&88&$B-S&'*#9:*8&0%&'-/0C%-K:'6&0/&
=,?
=,?T09:%*-2)J"$U
T09:%*-2)J"$U
+ ;*<3-=>?-6@'-.)%&/"'0&-A#5&8&$B-)8*"-0((&'-C&0D0$/B-.&%%&B-4'('&06-&%9E
Æ
Æ L)9:&-("$")
L)9:&-("$")7%0(#8#*-*+,-*.")#"/"0*)'U
7%0(#8#*-*+,-*.")#"/"0*)'U
Gruppe
P300
Schuldig
9/11 (82 %)
Unschuldig
1/11 (9 %)
‚Countermeasure‘
2/11 (18 %)
Der Tatwissentest
Ø Vorwiegend im Labor eingesetzt
Ø Theoretische Grundlage
Ø Potenzial und Limitationen des Tests abschätzbar
Ø G e g e n m a ß n a h m e n ( ‚ c o u n t e r m e a s u r e s ‘ ) s i n d
ungenügend erforscht
USING IMAGING TO IDENTIFY DECEIT: SCIENTIFIC AND ETHICAL QUESTIONS
Using Imaging
to Identify
Deceit
AMERICA
Scientific and
Ethical Questions
Quelle: American Academy of Sciences and Arts
Quelle: www.bion.de
Psychologie
Physiologie
ZNS
Struktur 1
Lüge
ZNS
Struktur 2
Cerebral Cortex July 2009;19:1557--1566
doi:10.1093/cercor/bhn189
Advance Access publication November 2, 2008
The Contributions of Prefrontal Cortex and
Executive Control to Deception: Evidence
from Activation Likelihood Estimate Metaanalyses
Shawn E. Christ1, David C. Van Essen2, Jason M. Watson3,
Lindsay E. Brubaker1 and Kathleen B. McDermott4
Previous neuroimaging studies have implicated the prefrontal
cortex (PFC) and nearby brain regions in deception. This is
consistent with the hypothesis that lying involves the executive
control system. To date, the nature of the contribution of different
aspects of executive control to deception, however, remains
unclear. In the present study, we utilized an activation likelihood
estimate (ALE) method of meta-analysis to quantitatively identify
brain regions that are consistently more active for deceptive
responses relative to truthful responses across past studies. We
then contrasted the results with additional ALE maps generated for
3 different aspects of executive control: working memory, inhibitory
control, and task switching. Deception-related regions in dorsolateral PFC and posterior parietal cortex were selectively associated
with working memory. Additional deception regions in ventrolateral
PFC, anterior insula, and anterior cingulate cortex were associated
with multiple aspects of executive control. In contrast, deceptionrelated regions in bilateral inferior parietal lobule were not
associated with any of the 3 executive control constructs. Our
findings support the notion that executive control processes,
particularly working memory, and their associated neural substrates play an integral role in deception. This work provides
a foundation for future research on the neurocognitive basis of
2005; Kozel et al. 2005; Nuñez et al. 2005; Phan et al. 2005; Abe
et al. 2006; Mohamed et al. 2006).
As with prior research, recent neuroimaging studies may be
conceptualized as arising from 1 of 2 primary motivations: 1) to
detect deception or 2) to differentiate of the neurocognitive
processes underlying deception (‘‘differentiation of deception,’’
Furedy et al. 1988). Although these aims are not necessarily
mutually exclusive, it is often the case that studies designed for
one purpose are not optimal for the other (for additional
discussion, see Furedy et al. 1988).
For example, in one of the first neuroimaging studies on
deception, Langleben et al. (2002) employed a variation of the
Guilty Knowledge Test (GKT) (Lykken 1959, 1960; Furedy and
Ben-Shakhar 1991), a questioning technique that has been used
extensively in the forensic field. Importantly, the traditional
GKT relies not only on the detection of deception per se but
also on the detection of recognition memory for details of the
crime scene (Ben-Shakhar and Elaad 2003). In the study of
Langleben et al., participants were given a playing card (e.g., a 5
of clubs) and instructed to deny their possession of the card
when later queried. Once in the scanner, participants were
1
Department of Psychological Sciences, University of Missouri,
Columbia, MO 65203, USA, 2Department of Anatomy and
Neurobiology, Washington University School of Medicine St.
Louis, MO 63110 USA, 3Department of Psychology, University
of Utah, Salt Lake City, UT 84112 USA and 4Department of
Psychology, Washington University in St Louis, St. Louis, MO
63130 USA
Downloaded from http://cercor.oxfordjournals.org/ at University
Meta-analysis
with the notion that deception is a more demanding
k than simply telling the truth, few studies report regions
ater activation for truthful responding as compared with
sponding (but see Langleben et al. 2005). Accordingly, we
present efforts on the much more common findings of
ural activation for deceptive relative to truthful responses.
ne with previous ALE meta-analyses (e.g., Turkeltaub et al.
aum et al. 2005; Owen et al. 2005), the measure of interest
ation of such activation rather than effect size (see
ordjournals.org/ at University of Konstanz, Library on May 11, 2012
d Methods
this latter construct.
The inclusion/exclusion criteria were generally similar across the
aforementioned studies and the present deception ALE meta-analysis
(e.g., use of PET or fMRI methodology, 3-dimensional [3D] coordinates
reported in stereotaxic space, inclusion of canonical contrast of
2 conditions, data from neurologically uncompromised participants).
Detailed descriptions of the inclusion criteria are available in the
original publications (Buchsbaum et al. 2005; Laird, McMillan, et al.
2005; Owen et al. 2005) and are not reproduced here. Any modest
variations in criteria across studies reflect the relative maturity and
current state of research in the respective area; therefore, these
differences were maintained for the present comparisons. For
example, in the case of the working memory and inhibitory control
ALE maps, there was sufficient data available to focus in on a specific
experimental paradigm (n-back task and Stroop task, respectively),
whereas for task switching and the present deception ALE maps, this
was not possible.
The Rationale:
„Lying is more effortful than telling the truth“
y appropriate articles for the deception meta-analysis,
e electronic databases (e.g., PsychInfo, MedLine, PubMed)
ed in April 2008 using various combinations of relevant
(e.g., deception, lying, fMRI, PET, MRI, neuroimaging). The
lusion/exclusion criteria were used to select articles for
meta-analysis:
cles that utilized PET or fMRI methodology were
d. Electrophysiological- (e.g., electroencephalography,
ncephalography, skin conductance response [SCR]) and
-only studies were excluded. Both blocked and eventudies were allowed, in order to obtain sufficient data to
he meta-analysis. Activations recorded using a block design
sent both transient item-related activity as well as sustained
ated to task set (Visscher et al. 2003). Activations observed
elated studies reflect only the transient component. As
n regions identified via the meta-analysis as showing
activation across these 2 different types of studies likely
e common component: item-related activity. Similar
prevented us from considering additional dimensions
monalities/differences in behavioral paradigms, sample
stics, etc.) in separate meta-analyses.
cles with experiments that yielded a clear contrast
ng locations of greater activation for deceptive responding
red with telling the truth and that did not include an
imitation (e.g., confound with recognition memory;
et al. 2002) were included.
Table 1
Data sources included in the deception meta-analysis
Publication
Modified GKT paradigm
Langleben et al. (2005)
Phan et al. (2005)
Past personal information/experience
Ganis et al. (2003)
Lee et al. (2002)—Experiment 2
Nuñez et al. (2005)
Spence et al. (2004)
Spence et al. (2008)
Recent action events
Abe et al. (2006)
Kozel et al. (2005)
Spence et al. (2001)
Recent knowledge
Kozel, Padgett, and George (2004)
Kozel, Revell, et al. (2004)
Lee et al. (2002)—Experiment 1
Foci
Method
Original
stereotaxic
space
Response
19
11
fMRI
fMRI
SPM2
SPM99
Manual
Manual
12
22
8
5
7
fMRI
fMRI
fMRI
fMRI
fMRI
AFNI
AFNIa
SPM2
SPM99
SPM2
Manual and verbal
Manual
Manual
Verbal
Verbal
4
32
6
PET
fMRI
fMRI
SPM2
SPM2
SPM99
Verbal
Manual
Manual
11
10
26
fMRI
fMRI
fMRI
SPM2
SPM96
AFNIa
Manual
Manual
Manual
a
The authors appear to have utilized a nonstandard registration algorithm; however, the target
space (T88) was identical to that employed in AFNI.
Cerebral Cortex July 2009, V 19 N 7 1559
Christ et al., 2009
BOLD-Aktivierungen
der Meta-Analyse
that they be separated by 12 mm, or else, the peaks were
consolidated by coordinate averaging. Regions around the peak
activations were identified by choosing contiguous voxels
within 10 mm of the peak activation that surpassed the
statistical threshold within the z plane of peak activity and in
Contributions of Different Paradigms
As noted earlier, sufficient data were not available to conduct
a separate meta-analysis for each paradigm type. However,
visual inspection of Figure 1 confirms that foci from multiple
paradigms appear to contribute to each of the ROIs. For
Downloaded from http://cercor.oxfordjournals.org/ at University of Kon
Figure 1. Previously reported foci demonstrating greater activation for deceptive responses (i.e., lies) as compared with truthful responses overlaid on the results from the ALE
meta-analysis. ALE data were thresholded at a value of 0.00502 (which corresponds to p \ 0.05 False Discovery Rate corrected). Foci were projected by study-specific
stereotaxic projection to the PALS-B12 atlas surface (see Materials and Methods) and are viewed on the inflated PALS atlas surface (Van Essen 2005), color coded based on
paradigm type/content. The upper and lower panels show foci in relation to lateral and medial views of the average fiducial surface, respectively. Selected classical Brodmann
areas (black borders) as well as orbitofrontal areas (light blue borders) from Öngür et al. (2003) are also illustrated. On all surfaces, foci are shown ‘‘pasted’’ to the surface,
irrespective of whether their 3D coordinates lie above or below the surface.
Table 2
ROIs identified from ALE analysis (deception [ truth) in FLIRT stereotaxic space
Region
Location
BA
1
2
3
4
5
6
7
8
9
10
11
12
13
Right insula
Right IFG
Right middle frontal gyrus
Right inferior parietal lobule/supramarginal gyrus
Right internal capsule/thalamus
Left IFG
Left inferior parietal lobule
Left internal capsule
Left insula
Left precentral gyrus/middle frontal gyrus
Right insula
Right anterior cingulate
Right inferior parietal lobule
NA
6/44/45
9/10/46
39/40
NA
44
40
NA
NA
6
NA
24/32
7/39
Peak activation
x
y
z
37
52
32
59
13
!49
!57
!16
!35
!42
35
5
47
20
14
43
!50
!5
15
!49
1
13
1
30
20
!65
!6
6
26
29
12
!4
31
13
2
53
!5
34
42
Volume (cm3)
ALE value 3 103*
3.7
3.7
3.4
2.0
1.5
1.4
1.2
0.8
0.9
0.4
0.9
0.8
0.3
10.09
10.13
9.29
8.33
8.13
7.23
7.12
6.41
6.36
6.11
6.65
5.54
5.64
BOLD-Aktivierungen
kognitiver Prozesse
Figure 2. Results of the working memory (green), inhibitory control (red), task switching (blue), and deception (black borders) ALE analyses viewed on the inflated PALS atlas
surface (Van Essen 2005).
uniquely associated inhibitory control relative to working
memory and task switching. In terms of the task switching
ALE map, a region in the left occipital cortex (BA 19) was
associated with task switching but not working memory or
inhibitory control.
executive control and deception ALE maps in bilateral inferior
parietal lobule (regions 4 and 7) and right IFG (region 2).
Discussion
Downloaded from http://cercor.oxfordjournals.org/ at University of Konstanz, Library on May 11, 2012
Note: NA, not applicable.
*P \ 0.05 (FDR corrected) in all instances.
24 x
24 x
Ja
Nein
Restliche
Karten
24 x
168 x
Nein
Nein
Langleben et al., 2005
Klassifikation – 4 neue Fälle
Klassifikation – 4 neue Fälle
Sensitivität
68.8 %
P(Lüge erkannt/tatsächlich Lüge)
Spezifität
83.7 %
P(Wahrheit erkannt/tatsächlich Wahrheit )
NeuroImage 55 (2011) 312–319
Contents lists available at ScienceDirect
NeuroImage
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y n i m g
Lying in the scanner: Covert countermeasures disrupt deception detection by
functional magnetic resonance imaging
Giorgio Ganis a,b,c,⁎, J. Peter Rosenfeld d, John Meixner d, Rogier A. Kievit e, Haline E. Schendan b,c
a
Department of Radiology, Harvard Medical School, Boston, MA 02115, USA
Massachusetts General Hospital, Martinos Center, Charlestown, MA 02129, USA
School of Psychology, University of Plymouth, Plymouth, Devon, PL48AA, UK
d
Department of Psychology, Northwestern University, Evanston, IL 60208-2710, USA
e
Department of Psychology, University of Amsterdam, Amsterdam, 1018WB, Netherlands
b
c
a r t i c l e
i n f o
Article history:
Received 7 September 2010
Revised 27 October 2010
Accepted 5 November 2010
Available online 24 November 2010
a b s t r a c t
Functional magnetic resonance imaging (fMRI) studies have documented differences between deceptive and
honest responses. Capitalizing on this research, companies marketing fMRI-based lie detection services have
been founded, generating methodological and ethical concerns in scientific and legal communities. Critically,
no fMRI study has examined directly the effect of countermeasures, methods used by prevaricators to defeat
deception detection procedures. An fMRI study was conducted to fill this research gap using a concealed
information paradigm in which participants were trained to use countermeasures. Robust group fMRI
differences between deceptive and honest responses were found without, but not with countermeasures.
Furthermore, in single participants, deception detection accuracy was 100% without countermeasures, using
activation in ventrolateral and medial prefrontal cortices, but fell to 33% with countermeasures. These
findings show that fMRI-based deception detection measures can be vulnerable to countermeasures, calling
for caution before applying these methods to real-world situations.
© 2010 Elsevier Inc. All rights reserved.
314
G. Ganis et al. / NeuroImage 55 (2011) 312–319
Fig. 1. Concealed information task paradigm. Schematic of stimuli employed in the no knowledge (NK), concealed knowledge (CK), and countermeasure (CM) conditions. Stimuli
included irrelevant dates and an infrequent probe date. Irrelevant dates were nonsalient dates with no particular meaning to participants. In the no knowledge condition, the probe
was an additional irrelevant date; hence, in this condition, participants had no knowledge about the probe date. In the concealed knowledge condition, the probe was the birth date
of each participant. There was also a third type of stimulus, an infrequent target date studied before the fMRI session, to ensure that participants had to attend the stimuli to perform
the task. Participants responded truthfully to all irrelevant and target dates (“no” and “yes”, respectively) and deceptively (“no”) to the probe date. The countermeasure condition
was the same as the concealed knowledge condition, but participants performed 3 distinct countermeasures on 3 of the irrelevant dates, just before indicating whether they knew
the dates.
Zielreiz
Geburtstag
and they were asked to cross out any dates they knew. All dates
described as salient earlier were also crossed out by participants
during this verification step.
In the main group, participants were tested in the no knowledge,
concealed knowledge, and countermeasure conditions. The no
knowledge condition was administered before the concealed knowledge condition because after performing this condition (which
required lying) participants would have become aware of the purpose
of the study and might have not been able to act as individuals with no
Countermeasure:
Eine nicht wahrnehmbare Bewegung mit...
CM 1 = linken Zeigefinder
placedCM
behind
of participants.
Participants saw the screen via
2the
= head
linkem
Mittelfinger
a front-surface mirror on the head coil.
CMwere
3 =analyzed
linken
großen
Images
with
AFNI (Cox,Zehen
1996) as follows: (a) slice
timing correction; (b) motion correction; (c) spatial smoothing with a
Gaussian filter (full-width half-maximum = 6 mm); (d) amplitude
normalization, by scaling timeseries to a mean of 100 and calculating
the percent signal change about this mean; (e) spatial normalization
to the MNI305 template; and (f) spatial resampling to a 3 × 3 × 3 mm
grid. For the hemodynamic response function, a gamma-variate
316
Concealed Knowledge Group
Probe > Irrelevant
G. Ganis et al. / NeuroImage 55 (2011) 312–319
Fig. 2. Differences between probes and the mean of irrelevant items in the main group (n = 12) for the concealed knowledge condition (p b 0.01, FDR corrected for multiple
comparisons), shown on an inflated brain (top: lateral view; bottom: medial view). The color scale depicts percent signal change. The seven activation clusters labeled and indicated
by yellow ellipses were found also using the same contrast in the ROI group. Note that the 3 medial regions were combined into single bilateral clusters. Abbreviations for the brain
region labels are as in Table 1.
concealed knowledge condition using block as factor (collapsing
across ROIs) showed this not to be the case, F(4,44) = 1.03, p N 0.1,
η2P = 0.094 (Block 1: M = 0.29, SE = 0.048; Block 2: M = 0.23,
SE = 0.053; Block 3: M = 0.25, SE = 0.048; Block 4: M = 0.30,
0.50, p N 0.1, η2P = 0.047 (Block 1: M = 0.044, SE = 0.026; Block 2:
M = 0.049, SE = 0.027; Block 3: M = 0.068, SE = 0.029; Block 4:
M = 0.050, SE = 0.028; Block 5: M = 0.048, SE = 0.030).
Moreover, using a 3-stimulus protocol of the type used here,
318
G. Ganis et al. / NeuroImage 55 (2011) 312–319
Fig. 4. Results of the single subject analyses: (a) bar graphs showing group activation to probes and irrelevants in the ROI triplet that best discriminated concealed and no knowledge
cases (100% accuracy and largest margin): left GFi/INS, right GFi/INS, and GC/GFs/GFd. Error bars denote the standard error of the mean. Note that, given the differential response to
targets required by the task, comparing brain activation between targets and irrelevants (or probes) is not informative. (b) Classification performance for all conditions achieved by
the classifier trained to discriminate no knowledge and concealed knowledge cases, assessed with a jackknife procedure using the best ROI triplet. Each data point is a test case; the
vertical axis shows the signed distance from the classification hyperplane, normalized by the maximum distance. Data were coded so that correctly classified no knowledge cases
would have a negative signed distance, whereas correctly classified concealed knowledge and countermeasure cases would have a positive signed distance. All 12 concealed and no
knowledge cases are classified correctly, but only 4 out of the 12 countermeasure cases are classified correctly.
methods to quickly associate new irrelevants with mental actions or
memories (e.g., via imagery); such associations could be established
during the first few trials and carried out consistently throughout
the test.
Given that these countermeasures can be learned easily, this
Acknowledgment
This research was supported in part by the National Science
Foundation (BCS0322611).
Lügendetektion und
funktionelle Kernspintomographie
Ø Laborstudien zeigen heterogene Resultate
Ø Geringe theoretische Grundlage
Ø Problem der Gegenmaßnahmen
... more science needed ...
Ø Integrative Theorie: Lüge, Emotion, Selbstregulation,
‚Mind reading‘, etc...
Ø Labor- und reale Studien
Ø Umfassende Betrachtung von ‚Countermeasures‘
Ø Integration multipler Paradigmen und physiologischer
Kenngrößen
Ø Neue Paradigmen und Methoden
Primary
Sensory
Cortex
Sensory
Thalamus
Emotional
Stimulus
Unimodal
Association
Cortex
Polymodal
Association
Cortex
Lateral Nucleus
Entorhinal
Cortex
medial
lateral
Basal
Amygdala
Accessory
Basal
PVN
Anterior
Pituitary
Stress
Hormones
Subiculum
Nucleus
Basalis
Lateral
Hypothalamus
Central
Gray
RPC
RVL
Medulla
DMV
NA
Parasympathetic
Activation
Cortical
Arousal,
Attention
Hippocampus
Central
BNST
Parabrachial
ventrolateral
Emotional
Behavior
Sympathetic
Activation
Startle
Reflex
Potentiation
Davis, 1992, 1994
LeDoux, 2000
Primary
Sensory
Cortex
Unimodal
Association
Cortex
Lateral Nucleus
Sensory
Thalamus
Emotional
Stimulus
medial
lateral
Basal
Amygdala
Parabrachial
PVN
Stress
Hormones
Entorhinal
Cortex
Cortical
Arousal,
Attention
Hippocampus
Central
Subiculum
Nucleus
Basalis
Lateral
Hypothalamus
Central
Gray
RPC
RVL
Medulla
DMV
NA
Parasympathetic
Activation
ventrolateral
Accessory
Basal
BNST
Anterior
Pituitary
Polymodal
Association
Cortex
Emotional
Behavior
Sympathetic
Activation
Startle
Reflex
Potentiation
Primary
Sensory
Cortex
Unimodal
Association
Cortex
Lateral Nucleus
Sensory
Thalamus
Emotional
Stimulus
medial
lateral
Basal
Amygdala
Parabrachial
PVN
Stress
Hormones
Entorhinal
Cortex
Cortical
Arousal,
Attention
Hippocampus
Central
Subiculum
Nucleus
Basalis
Lateral
Hypothalamus
Central
Gray
RPC
RVL
Medulla
DMV
NA
Parasympathetic
Activation
ventrolateral
Accessory
Basal
BNST
Anterior
Pituitary
Polymodal
Association
Cortex
Emotional
Behavior
Sympathetic
Activation
Startle
Reflex
Potentiation
EKP-Differenz: Erotischer – neutraler Kontext
1.5
0
µV
-1.5
Primary
Sensory
Cortex
Unimodal
Association
Cortex
Lateral Nucleus
Sensory
Thalamus
Emotional
Stimulus
medial
lateral
Basal
Amygdala
Parabrachial
PVN
Stress
Hormones
B (N)
Entorhinal
Cortex
Cortical
Arousal,
Attention
Hippocampus
Subiculum
Central
Nucleus
Basalis
Lateral
Hypothalamus
Central
Gray
RPC
RVL
Medulla
DMV
NA
Parasympathetic
Activation
ventrolateral
Accessory
Basal
BNST
Anterior
Pituitary
Polymodal
Association
Cortex
Emotional
Behavior
Sympathetic
Activation
Startle
Reflex
Potentiation
U (?)
B (N)
B (E)
U (?)
U (?)
Primary
Sensory
Cortex
Unimodal
Association
Cortex
Lateral Nucleus
Sensory
Thalamus
Emotional
Stimulus
medial
lateral
Basal
Amygdala
Parabrachial
PVN
Stress
Hormones
Entorhinal
Cortex
Cortical
Arousal,
Attention
Hippocampus
Central
Subiculum
Nucleus
Basalis
Lateral
Hypothalamus
Central
Gray
RPC
RVL
Medulla
DMV
NA
Parasympathetic
Activation
ventrolateral
Accessory
Basal
BNST
Anterior
Pituitary
Polymodal
Association
Cortex
Emotional
Behavior
Sympathetic
Activation
Startle
Reflex
Potentiation
Bekannte Personen: Erotischer – neutraler Kontext
1
0
-1
µV
Topics focusing on the potential use of brain-based (especially fMRI-based) lie detection have attracted a great
deal of attention. There are several published reviews summarizing the results of recent studies of fMRI lie detection and discussing its feasibility (Bles and Haynes 2008; Haynes 2008; Sip and others 2008). As pointed out by
many researchers, the results obtained in highly controlled laboratory settings are qualitatively different from
those obtained in real-life situations. In addition, the effects of simple countermeasure (a method used by liars
to defeat lie-detection procedures) on fMRI lie detection remain unclear. Some ethical issues should be noted,
such as privacy concerns. At the present time, brain-based lie detection is an imperfect technology, and a cautionary stance is necessary.
Lügendetektor -­‐ nur ein Mythos? Downloaded from nro.sagepub.com at Universitaet Konstanz on May 7, 2012
Quelle: Abe, 2011